Cadence at Yotta 2026: From Infrastructure Investment to AI Value

The AI industry has entered a new phase. The conversation is no longer all about acquiring the latest GPUs. Instead, organizations are focused on a more fundamental question: How quickly can we convert power, infrastructure, and capital into productive AI output?

That challenge is at the heart of Yotta 2026, where the industry's leading hyperscalers, AI-cloud providers, colocation operators, technology vendors, investors, utilities, and infrastructure developers will come together to address the future of AI infrastructure. Industry materials describe Yotta as an event focused on the intersection of AI, digital infrastructure, energy, cooling, deployment, and operations, reflecting the growing recognition that AI infrastructure is a system-wide challenge rather than a collection of independent technologies.

Cadence is excited to be part of that conversation.

AI Factories Change the Rules

AI factories have emerged as a new class of digital infrastructure. Their success is measured not simply by capacity or uptime, but by the ability to produce tokens quickly, efficiently, and profitably.

As operators scale to increasingly dense GPU deployments, traditional design approaches are being challenged by interconnected constraints across power delivery, liquid cooling, networking, facility infrastructure, and operations. Small inefficiencies can translate into significant impacts on deployment schedules, operational costs, utilization, and ultimately revenue generation.

The industry's focus is rapidly shifting from infrastructure planning to infrastructure optimization. Organizations are increasingly evaluating metrics such as utilization, cooling efficiency, resiliency, and tokens generated per watt of energy consumed.

This is where digital twins are becoming essential.

Join Cadence for Our Executive Lunch Session

The Economics of AI Factories: Accelerating Time-to-Token and Maximizing Capital Efficiency

Tuesday, September 29

Doors Open: 12:50pm | Program: 1:00–2:00pm

CAESARS FORUM | Las Vegas

Registration Required | Seating Limited to 100 Attendees

Featured Speakers

Join industry leaders and experts for a panel discussion featuring:

• Vladimir Troy – Vice President of AI Infrastructure, NVIDIA
• David Quirk – President and CEO, DLB Associates
• Fred Rebarber – Vice President, Technical Business Development Neocloud, Vertiv
• Moderator: Sherman Ikemoto – AI and Data Center Group Director, Cadence

During this exclusive lunch session, panelists will explore how organizations can reduce deployment risk and accelerate the journey from infrastructure investment to AI revenue.

The discussion will examine the challenges of designing, deploying, and operating next-generation AI infrastructure, and how digital twins and advanced simulation can help organizations identify infrastructure constraints before they become costly operational problems. Attendees will gain practical insights into the relationship between infrastructure performance, token production, profitability, and return on investment.

Whether you're building a hyperscale AI factory, expanding a colocation footprint, or deploying sovereign AI infrastructure, understanding infrastructure behavior before physical deployment is becoming a critical competitive advantage.

Visit the Cadence Booth

At the Cadence booth 1019, attendees can see firsthand how the Cadence Reality Digital Twin Platform, accelerated by NVIDIA Omniverse libraries and built on OpenUSD, supports the entire AI-factory lifecycle, from design and deployment through operations.

Demo: AI Factory and Data Center Design

Designing an AI factory today requires optimization across multiple engineering domains simultaneously. Power systems, cooling infrastructure, airflow, liquid networks, rack layouts, and facility constraints must operate as a unified system.

Leveraging the Cadence Reality Digital Twin Platform, organizations can rapidly create high-fidelity digital representations of AI factories using the extensive Cadence Reality DC Elements library, including models such as the NVIDIA GB300 NVL72 platform SimReady asset. Integrated with the NVIDIA Omniverse DSX Blueprint, part of NVIDIA DSX Sim, these models support thermal and cooling validation before deployment. These validated infrastructure components accelerate model creation and enable faster evaluation, optimization, and deployment of next-generation AI infrastructure.

This demo showcases how teams can simulate and assess infrastructure performance before deployment, helping identify bottlenecks, validate designs, and reduce costly rework. By combining physics-based simulation, digital twins, and pre-configured AI infrastructure models, organizations can make more informed decisions, reduce deployment risk, and accelerate time-to-first token.

Demo: AI Factory and Data Center Operations

The challenge doesn't end once the facility is operational.

As AI factories scale, optimizing individual components is no longer enough. Operators must understand how compute, power, cooling, and facility infrastructure work together as a unified system to deliver the desired outcome: maximum token production from every available watt and dollar invested. Operational digital twins provide a virtual environment where teams can evaluate changes, predict cross-domain impacts, and optimize system-level performance before modifying live systems.

This demo showcases how organizations can combine real-world operational telemetry, physics-based simulation, and AI surrogate models to move from reactive monitoring to predictive optimization. Trained on high-fidelity simulations, AI surrogates enable near real-time evaluation of infrastructure scenarios, allowing operators to explore workload placement options, identify capacity constraints, and optimize power and cooling strategies in minutes rather than hours. By combining rapid AI-driven insights with engineering-grade validation, organizations can improve decision-making, maximize infrastructure performance, and increase AI productivity throughout the facility lifecycle.

From Power to Tokens

One of the biggest themes emerging across the AI infrastructure industry is the recognition that power has become the primary constraint. Organizations are increasingly focused on maximizing the value generated from every available megawatt while accelerating deployment timelines and improving operational efficiency.

The Cadence Reality Digital Twin Platform helps address these challenges by connecting design, simulation, and operational intelligence into a unified digital thread. By enabling teams to model, predict, and optimize infrastructure behavior across the entire lifecycle, Cadence helps organizations reduce risk, improve performance, and accelerate time-to-first token with better token-per-watt output.

If you're attending Yotta 2026, join us for our lunch session and stop by the booth to learn how digital twins can help transform AI infrastructure from a collection of components into a high-performing AI factory.